klbr/n8nworkflows.xyz

Fork 0

mirror of https://github.com/khoaliber/n8nworkflows.xyz.git synced 2026-04-22 17:09:32 +00:00

Files

T

nusquama e55698f16d creation

2025-11-12 16:08:43 +01:00

15 KiB

Raw Blame History

Receipt Scanning & Analysis Workflow

https://n8nworkflows.xyz/workflows/receipt-scanning---analysis-workflow-7990

Receipt Scanning & Analysis Workflow

1. Workflow Overview

This workflow automates the processing of newly added invoice files in a specific Google Drive folder. It extracts text from scanned invoice PDFs, analyzes their content using AI models, structures the extracted data, and appends the relevant invoice details into a Google Sheets document for record-keeping and further processing.

The workflow is composed of the following logical blocks:

1.1 Input Reception (File Detection and Download): Detects newly created files in a designated Google Drive folder and downloads them.
1.2 Text Extraction: Uses an AI-powered OCR model to extract text content from the downloaded invoice PDFs.
1.3 AI Invoice Analysis: Applies an AI agent with embedded OpenAI and structured output parsers to analyze and interpret the invoice text.
1.4 Data Storage: Appends the structured invoice data into a Google Sheets spreadsheet for tracking.
1.5 Auxiliary: Includes a sticky note with tutorial information unrelated to main logic.

2. Block-by-Block Analysis

2.1 Input Reception (File Detection and Download)

Overview:
Watches a specific Google Drive folder for newly created files (invoices) and downloads each detected file for further processing.
Nodes Involved:
- Google Drive Trigger
- HTTP Request
Node Details:
- Google Drive Trigger
  - Type: Trigger node, listens for Google Drive events
  - Configuration: Watches the folder with ID 12aSoGo009rsrhyJ_a4VP84m_NrWCsA3O (named "Invoices") for new files created (fileCreated event), polling every minute
  - Credentials: Google Drive OAuth2
  - Inputs: None (trigger)
  - Outputs: Emits metadata of the newly created file(s)
  - Edge Cases:
    - Folder access permissions issues
    - Polling delay impacts near real-time detection
    - Large file creation events batching
    - Token expiration or revoked credentials causing trigger failure
- HTTP Request
  - Type: HTTP node to download file content
  - Configuration:
    - Constructs URL dynamically using file ID from trigger (https://www.googleapis.com/drive/v3/files/{{ $json.id }}?alt=media)
    - Response format set to download the file as a binary (responseFormat: file)
    - Uses Google Drive OAuth2 credentials (different account from trigger, ensure consistent permissions)
  - Inputs: Receives file metadata from Google Drive Trigger
  - Outputs: Binary file content of the invoice PDF
  - Edge Cases:
    - File deletion between detection and download
    - API rate limits or timeouts
    - Credential mismatches or insufficient permissions
    - Large file size causing memory or timeout issues
  - Version Notes: Node version 4.2 supports improved binary handling

2.2 Text Extraction

Overview:
Extracts textual content from the downloaded invoice PDF using an AI-powered OCR or text extraction model.
Nodes Involved:
- Extract text
Node Details:
- Extract text
  - Type: Mistral AI node (AI text extraction/OCR)
  - Configuration: Default options (no custom parameters specified)
  - Credentials: Mistral Cloud API
  - Inputs: Binary PDF file from HTTP Request node
  - Outputs: JSON object containing extracted text, including page-wise markdown representation under pages[0].markdown
  - Edge Cases:
    - Poor quality scans causing inaccurate text extraction
    - Unsupported file format or corrupted PDF
    - API errors or authentication failures
    - Latency or throttling on Mistral API side

2.3 AI Invoice Analysis

Overview:
Uses an AI agent incorporating OpenAI GPT-4 and a structured output parser to analyze the extracted invoice text and produce a structured JSON with invoice details.
Nodes Involved:
- AI Agent
- OpenAI Chat Model
- Structured Output Parser
Node Details:
- AI Agent
  - Type: Langchain AI agent node
  - Configuration:
    - Prompt: Instructs the AI as an invoice analysis expert and injects the extracted markdown text from the first page of the invoice ({{ $json.pages[0].markdown }})
    - Prompt type: Define (custom prompt)
    - Output parser enabled to parse AI output into structured JSON
  - Inputs: Extracted text JSON from Mistral AI node
  - Outputs: Structured JSON with invoice fields
  - Edge Cases:
    - Prompt formatting or expression errors
    - Unexpected or incomplete text input causing misinterpretation
    - AI model unavailability or latency
- OpenAI Chat Model
  - Type: Langchain OpenAI chat completions node
  - Configuration:
    - Model: GPT-4.1-mini (a variant of GPT-4 optimized for smaller size)
    - No additional options specified
  - Credentials: OpenAI API key
  - Inputs: Connected as AI language model for the AI Agent node
  - Outputs: Receives AI completions for the agent's prompt
  - Edge Cases:
    - API rate limits or quota exhaustion
    - Model version deprecation or changes
    - Network or authentication failures
- Structured Output Parser
  - Type: Langchain structured output parser node
  - Configuration:
    - Provides an example JSON schema for invoice data fields: invoiceNumber, From, To, invoiceDate, dueDate, shortDescription
    - Ensures AI output conforms to this structure for downstream use
  - Inputs: Connected as AI output parser for the AI Agent node
  - Outputs: Parsed structured data used by next node
  - Edge Cases:
    - Parsing errors if AI output does not match schema
    - Schema changes requiring updates in parser config

2.4 Data Storage

Overview:
Appends the structured invoice data into a Google Sheet to maintain a log of processed invoices.
Nodes Involved:
- Append row in sheet
Node Details:
- Append row in sheet
  - Type: Google Sheets node
  - Configuration:
    - Operation: Append
    - Target Sheet: Sheet1 (gid=0) of spreadsheet ID 14tyy2at8ZtfoQq4zGYph4DG7omv9uA9aXGH5KnQbPgs
    - Columns mapped: Invoice Number, From, To, Invoice Date, Due Date, Short Description, Link to File
    - Values sourced from AI Agent node output fields and Google Drive Trigger's file link ($('Google Drive Trigger').item.json.webViewLink)
    - No type conversion or matching columns enabled
  - Credentials: Google Sheets OAuth2
  - Inputs: Structured invoice data JSON from AI Agent node
  - Outputs: Confirmation of row append
  - Edge Cases:
    - Sheet access/permission issues
    - Data type mismatches or missing fields
    - Spreadsheet quota or API limits
    - Network errors or credential expiration

2.5 Auxiliary

Overview:
Provides tutorial link and author information as a sticky note for user reference.
Nodes Involved:
- Sticky Note

Node Details:

Sticky Note

Type: Visual annotation node, no runtime impact

Content:

## n8n Starter Session Tutorial

[Watch the Tutorials](https://www.youtube.com/playlist?list=PLWYu7XaUG3XOJwOOGiX89SQ_w67vw3dq7)

Tutor: [Aemal Sayer](https://aemalsayer.com)

Position: Top-left area, separate from main flow
Edge Cases: None

3. Summary Table

Node Name	Node Type	Functional Role	Input Node(s)	Output Node(s)	Sticky Note
Google Drive Trigger	Google Drive Trigger (trigger)	Detect new invoice files	None	HTTP Request
HTTP Request	HTTP Request (file download)	Download invoice PDF	Google Drive Trigger	Extract text
Extract text	Mistral AI (OCR/text extraction)	Extract text from PDF	HTTP Request	AI Agent
AI Agent	Langchain AI Agent	Analyze invoice text	Extract text, OpenAI Chat Model, Structured Output Parser	Append row in sheet
OpenAI Chat Model	Langchain OpenAI Chat Completion	Provide AI language model	None (invoked by AI Agent)	AI Agent
Structured Output Parser	Langchain Structured Output Parser	Parse AI output	None (invoked by AI Agent)	AI Agent
Append row in sheet	Google Sheets Append Row	Store parsed invoice data	AI Agent	None
Sticky Note	Sticky Note (annotation)	Tutorial link and credits	None	None	## n8n Starter Session Tutorial. Watch the Tutorials. Tutor: Aemal Sayer

4. Reproducing the Workflow from Scratch

Create Google Drive Trigger node
- Type: Google Drive Trigger
- Set event to fileCreated
- Set triggerOn to specificFolder
- Specify the folder ID to watch: 12aSoGo009rsrhyJ_a4VP84m_NrWCsA3O (your invoices folder)
- Set polling interval to every 1 minute
- Add Google Drive OAuth2 credential with appropriate permissions
Create HTTP Request node
- Type: HTTP Request
- Connect input from Google Drive Trigger node
- Set URL to: https://www.googleapis.com/drive/v3/files/{{ $json.id }}?alt=media (use expression to insert file ID dynamically)
- Under options → response, set responseFormat to file to download binary content
- Authentication: Use Google Drive OAuth2 credentials (ensure same or compatible account as trigger)
- Node version: Prefer version 4.2+ for improved binary support
Create Extract text node
- Type: Mistral AI (or equivalent OCR/text extraction AI node)
- Connect input from HTTP Request node (binary PDF)
- Use default options or configure as needed for your OCR model
- Attach Mistral Cloud API credentials
Create OpenAI Chat Model node
- Type: Langchain OpenAI Chat Model
- Select model: gpt-4.1-mini or suitable GPT-4 variant
- Attach OpenAI API credentials (API key with chat completion access)

Create Structured Output Parser node

Type: Langchain Structured Output Parser

Provide JSON schema example for invoice extraction fields:

{
  "invoiceNumber": "INV123",
  "From": "John street 23, 12443 Berlin, Germany",
  "To": "Smith street 33, 23222 Berlin, Germany",
  "invoiceDate": "20/08/2025",
  "dueDate": "25/08/2025",
  "shortDescription": "It is an invoice about web development costs"
}

No additional parameters needed

Create AI Agent node
- Type: Langchain AI Agent
- Set prompt type to define
- Insert prompt text:
```
You are an expert in analyzing invoices.

Here is an invoice that you should analyze: {{ $json.pages[0].markdown }}
```
- Enable output parser and select the structured output parser node
- Connect AI language model input to the OpenAI Chat Model node
- Connect AI output parser input to the Structured Output Parser node
- Connect input from Extract text node (which delivers OCR output)
Create Append row in sheet node
- Type: Google Sheets
- Set operation to append
- Configure document ID to target your invoice tracking spreadsheet (e.g., 14tyy2at8ZtfoQq4zGYph4DG7omv9uA9aXGH5KnQbPgs)
- Set the sheet name or gid (e.g., gid=0)
- Define columns and map them to AI Agent output fields:
  - Invoice Number → {{ $json.output.invoiceNumber }}
  - From → {{ $json.output.From }}
  - To → {{ $json.output.To }}
  - Invoice Date → {{ $json.output.invoiceDate }}
  - Due Date → {{ $json.output.dueDate }}
  - Short Description → {{ $json.output.shortDescription }}
  - Link to File → {{ $('Google Drive Trigger').item.json.webViewLink }}
- Attach Google Sheets OAuth2 credentials
Connect the nodes in order:
- Google Drive Trigger → HTTP Request → Extract text → AI Agent → Append row in sheet
- OpenAI Chat Model and Structured Output Parser nodes connect to AI Agent as language model and output parser respectively
(Optional) Create Sticky Note node
- Add for documentation or reference purposes with tutorial links and credits

5. General Notes & Resources

Note Content	Context or Link
n8n Starter Session Tutorial with step-by-step videos	YouTube Playlist
Tutorial author and expert contact	Aemal Sayer

Disclaimer:
The text provided originates exclusively from an automated workflow created with n8n, an integration and automation tool. This processing strictly complies with applicable content policies and contains no illegal, offensive, or protected elements. All manipulated data are legal and public.

15 KiB Raw Blame History

Receipt Scanning & Analysis Workflow

1. Workflow Overview

2. Block-by-Block Analysis

2.1 Input Reception (File Detection and Download)

2.2 Text Extraction

2.3 AI Invoice Analysis

2.4 Data Storage

2.5 Auxiliary

3. Summary Table

4. Reproducing the Workflow from Scratch

5. General Notes & Resources

15 KiB

Raw Blame History